Introduction to the Haskell Language

Functional programming model Functional Programming and supporting languages ​​appeared very early as soon as the concept of Declarative Programming appeared in the field of software programming in general. However, due to its heavy load of features borrowed from Mathematics, it is somewhat less friendly than Procedural Programming in terms of accessibility to potential programmers who are not from academic environments. Therefore, programming languages ​​based entirely on Functional Programming are often less popular and almost unknown to the majority of programmers before 2010.

However, that does not mean that Functional Programming is more difficult than Procedural Programming , or requires users to have a strong Math background. And even with a friendly self-study guide, with a slight slight on borrowed concepts from Mathematics, anyone can begin to learn, use, and see the effect that their way of thinking. Functional Programming provides.

Therefore, in the years after 2010, a series of popular programming languages ​​such as JavaScript (ECMA 2015) , Java 8 (Java 2014) , etc. have brought in new updated versions of Functional features to Help programmers write more concise, beautiful, and intuitive code. At the present time, perhaps no one is unaware Funtional Programming and a few related coding techniques such as lambda expressions, curry parameters, …

Haskell Language

In the group of purely Functional languages, there are two languages ​​that are considered to represent two schools of using different syntax to express Functional programming thinking, namely Haskell and LISP . Among them, Haskell is said to have a more user-friendly syntax for programmers who are familiar with popular programming languages. The reason is because the syntax of LISP is designed to be closer to the logic of computer reading, and here is an example of the function that calculates the value of the n element in the Fibonaci sequence being placed on the LISP homepage:

We’ll soon be rewriting this fib function ourselves using Haskell ‘s syntax, but first a short excerpt of the language’s history. Haskell was first introduced in 1990, several years earlier than the Java and JavaScript brothers (1995). This pure Functional language was designed by a team of programming engineers of nearly two dozen people and is named after the American mathematician and scientist Haskell Brooks Curry . This Sir is very well known in the academic environment at home and in addition to Haskell , the words Brooks and Curry are also used to name other programming languages, most of which are used in academic training and applications. industrial use.

And so, until about the years after 2010, the programming community in the world in general could not ignore the Funtional Programming that made Haskell become a highly influential language. Specifically, special module like LINQ/C# and Generics/Java are all designed with syntax and referential thinking from Haskell . Some other languages ​​that have received influence from Haskell include Scala , Swift , Visual Basic 9.0 , C++ 11 , F# , Rust , Elm , PureScript , etc..

Among them, Elm and PureScript are typical names that are also designed purely Functional with syntax similar to Haskell about ~ 90% . The rest of the difference is because of the features that Haskell supports and Elm and PureScript do not at the moment.

Hello, Haskell!

There are so many great things about Haskell and its development environment that we wouldn’t be able to list them all in one article. So it’s best to start right away with the Hello World code and the filename Main.hs – which can be created with MS Notepad or any other code editor.

You can test the starter code in the online code execution environment here: REPLit.com . We’ll talk about installing the Haskell -specific development environment in the next section of the article.

We have a Hello, World program that starts at main function which is defined by a data type declaration line with no input parameters and returns a value of type IO () , and a line describing The detailed operating logic is that main will delegate work to the putStrLn function, which is already defined in the Haskell standard library. Compared to Elm , the type declaration operation in Haskell is only different in that the :: notation is used to separate the function name and the type information of the parameters and return values. In Elm we only use 01 colon : .

The putStrLn function requires as an input a String string stored in the message variable, and the determination of the printed string is further delegated to the self-defined hello function with the input parameter name and type String and the return result. about is the body of the full greeting which is also of type String .

So the final result that we get after running the main function is a value of type IO () that has been defined by Haskell . As for how the logic will convert this value into a logic that makes the state change of the Console window, it is already implemented in the compiler and we don’t need to care.

Thus, the basic syntax of Haskell is not strange if you have accompanied me in the Sub-Series Elm . The main function is supported by other functions in the form of a computed expression and returns some result, rather than a description of the steps to be taken. A program written in Haskell will only be filled with function calls in expressions like this, and so function calls are designed to look pretty natural – without the parentheses like C or JavaScript .

Instead of writing pure Imperative statements describing the operation of calculating the string content to be printed first, and then running the printf procedure to print the content to the Console window; Then in Haskell we write code in the form of Declarative declaration to define main briefly as putStrLn message , and then continue to provide a detailed definition of the mesage in the next line with the keyword where is message = hello "Haskell" .

This is something Elm doesn’t have, and Elm only supports the let .. in construct which is also borrowed from Haskell but will make reading the code harder to follow in a Declarative mindset, because we’ll have to read the definition. Details about the message first, then read the rest of the main function. If we have many attached elements similar to the message , then when we want to skim through and extract the most concise definition information, main = text message to check the data type declaration of the main function, we will have to Specifies the last line in the let .. in in block to extract information describing the right side of the regex.

Of course, we can also stack function calls by pushing the priority of function calls behind with parentheses () like in other popular programming languages ​​instead of using the construct binding with where or let .. in .

Now the call to the hello function is written after putStrLn but will be executed first due to the precedence of the parentheses () surrounding the passed parameter. And we can see the first feature of Functional Programming that sub-program will always return a certain value, not directly creating side-effect on environmental elements. outside school.

Even the built-in putStrLn function will return a value of type IO () and can be passed to other function calls to perform calculations if necessary. The side-effect logic will be implemented automatically by the compiler similar to Elm which will return a data structure describing a single page SPA , but about the generated HTML and JS files like anyway will be optimized at the compiler level and the coder will not need to care about designing the procedure that does the job.

That’s why using a purely Functional Programming language in a specifically designed environment will have a vastly different experience than borrowing from Functional tools and using it in suboptimal Imperative environments. for Functional Programming .

This is similar to writing Procedural code in Ada which is much more convenient than borrowing tools like contract and then bringing it to popular languages ​​that are not fully optimized for Procedural Programming . Tools borrowed into environments that are not optimally designed for a particular programming paradigm will always have limitations in terms of functionality and readability of the syntax used.

Haskell Setup

To start installing the Haskell development toolkit on any operating system, simply open the official Haskell website at: Haskell.org -> GHC-up . The command to download and activate the GHC-up installer for the operating system you are using will appear right in the To install on ... panel of the website. All we need to do is copy this installation command and enter it into the command line window of the operating system we are using.

With Windows it would be Powershell :

As for Linux , it will be Terminal :

After that, the installation process will bring up some menus asking about the tools you want to install, so we just install all of them including the Stack project manager and HLS - Haskell Language Server . For example, if you want to install additional plug-in for VS Code to support syntax recognition and proofing in Haskell code files, the HLS will definitely be needed. For Windows , the installer will suggest installing the MSYS2 to use bash commands like Linux and installing more is not cumbersome. Just install it fully because there will be a time when you will need to use it.

After the installation is complete, we can type the command to check the version of the GHC compiler to confirm that the installation was successful.

And then proceed to compile and run the code of the main.hs file created in the previous example.